This study uses self-concept theory to examine the temporal stability in older adults' self-representations as one facet of their personality. Specifically, the proposed research will examine the stability of the context-independent self and the stability of role-specific selves in a sample of 160 older adults (young-olds: n=80; old-olds: n=80), at two times of measurement 3 months apart. This study extends previous research on older adults' self-representations in two ways: (1) it examines multiple self-facets and uses multiple methods of assessment; (2) it examines structural equation models of self-concept differentiation (SCD) and psychological outcomes, such as subjective well-being, depression, and anxiety. The specific aims of the study are: (1) To examine the stability of multiple self-representations in older adults over a 3-month period. It is expected that the context- independent self-representation will show significantly greater temporal stability than the role-specific self-representations. It is also expected that old-old adults show greater stability than young-old adults. (2) To examine the stability of the structural arrangement of older adults' self-representations. This study will examine whether there are differences in stability depending on the structural importance of self-descriptors. It is expected that self-descriptors of central and peripheral importance show greater stability than self- descriptors of intermediate importance. It is also hypothesized that this effect is constant across age groups. (3) To examine the convergent validity of three different indexes of SCD. Confirmatory factor analysis (CFA) will be used to examine the convergent validity of 3 separate indexes of SCD. It is expected that findings from the CFA will support the convergent validity of the three methods used. (4) To examine whether different indexes of SCD show the same relations with psychological outcomes. This study will examine whether the different indexes of SCD show the same predictive relations with measures of psychological functioning and personality, after the effects of age, education, and physical health have been controlled for. (5) To test conceptual models examining the effect of SCD on psychological outcomes. Structural equation modeling (SEM) will be used to examine direct and indirect predictor models of SCD and psychological outcomes. A third model will examine whether SCD itself is a mediator of the effects of other variables (e.g., self-esteem) on psychological outcomes.